Mapping genes for calcium signaling and their associated human genetic disorders.

MOTIVATION: Signal transduction via calcium ions (Ca2+) represents a fundamental signaling pathway in all eukaryotic cells. A large portion of the human genome encodes proteins used to assemble signaling systems that can transduce signals with diverse spatial and temporal dynamics.
RESULTS: Here, we provide a map of all of the genes involved in Ca2+ signaling and link these genes to human genetic disorders. Using Gene Ontology terms and genome databases, 1805 genes were identified as regulators or targets of intracellular Ca2+ signals. Associating these 1805 genes with human genetic disorders uncovered 1470 diseases with mutated 'Ca2+ genes'. A network with scale-free properties appeared when the Ca2+ genes were mapped to their associated genetic disorders.
AVAILABILITY AND IMPLEMENTATION: The Ca2+ genome database is freely available at http://cagedb.uhlenlab.org and will foster studies of gene functions and genetic disorders associated with Ca2+ signaling. CONTACT: per.uhlen@ki.se. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

1 Introduction

The calcium ion (Ca2+) is a universal cell-signaling messenger with diverse roles in a wide range of biological processes and human diseases (Berridge, 2012; Clapham, 2007; Uhlen and Fritz, 2010). A large number of Ca2+-channels, Ca2+-pumps and organelles are responsible for the strict regulation of intracellular Ca2+ concentrations (Fig. 1A) (Berridge ; Bootman ; Uhlen and Fritz, 2010) because high Ca2+ levels are toxic to the cell (Zhivotovsky and Orrenius, 2011). To date, more than 170 000 publications in NCBI's PubMed include Ca2+ in the title, thereby demonstrating the diverse role of Ca2+ in biological processes and the immense interest among scientists to understand the physiological and pathological processes regulated by Ca2+. However, this large quantity of available literature also makes it difficult to establish connections between disparate discoveries made over previous decades using diverse model systems.

Fig. 1

Regulating the cytosolic Ca2+ level. (A) Schematic diagram illustrating the main regulators of the cytosolic Ca2+ level. Cytosolic Ca2+ signals are generated through the concerted action of cellular mechanisms that increase (ON-reactions, red) and decrease (OFF-reactions, blue) the concentration of Ca2+ in the cytoplasm. Typical Ca2+ signals are initiated by stimuli that trigger the entry of external Ca2+ through ligand-gated channel receptors (LGCR), receptor operated channels (ROC) or voltage (ΔV) gated Ca2+ channels in the plasma membrane or by the activation of metabotropic receptors (mR) that stimulate PLC- and InsP3-mediated Ca2+-release from the ER/SR and the accompanying refilling of Ca2+ stores via store operated channels (SOC). When the cytosolic level of Ca2+ increases, Ca2+ itself stimulates InsP3Rs and/or RyRs to release further Ca2+ into the cytoplasm. During this phase, Ca2+ buffers bind Ca2+, which contributes to the decrease in the cytosolic concentration of free Ca2+. When the Ca2+ concentration reaches high levels, the plasma membrane Ca2+-ATPase (PMCA) and Na+/Ca2+-exchanger (NCX) extrude Ca2+ to the outside, whereas the ER/SR Ca2+-ATPase (SERCA) pumps Ca2+ back into the ER/SR. High levels of Ca2+ is also regulated by the mitochondria via the mitochondrial NCX (MNCX) and the Ca2+ uniporter. Additionally second messengers (2nd) and regulatory proteins (E) are able to modulate intracellular Ca2+. (B) Bar chart of 20 annotated GO terms associated with the most Ca2+ genes

Dysfunctional Ca2+ signaling has been reported in a large variety of human genetic disorders (Bidaud ; Chakraborti ; Delmas ; Giacomello ; Nedergaard ; Rajakulendran ; Ramasamy, 2008; Uhlen ). Given that Ca2+ regulates a myriad of physiological systems at every level, including the modulation of the cell’s membrane potential and underlying channels and ion transporters, kinases and transcription factors, signaling cascades that regulate gene expression, as well as tissue network activity via gap junctions and peptide/transmitter release to name a few, it should not be surprising that disruptions of Ca2+ homeostasis both extra- and intracellularly, underlie a host of diseases in virtually every organ system in the human body. For example, Ca2+ signaling plays a decisive role in the blood coagulation pathway (Triplett, 2000), a critical role in muscle contraction (smooth, striated and cardiac) with many skeletal myopathies resulting from intracellular Ca2+ dysregulation (Hernandez-Ochoa ), and clear links in Ca2+-handling gene mutations to inherited arrhythmia syndromes (Burashnikov ; Priori ). Moreover, autosomal dominant polycystic kidney disease, the most common inherited pathology of the kidneys, is caused by the mutation of two genes (PKD1 and PKD2) that encode integral membrane proteins that can transport Ca2+ (Kuo and Ehrlich, 2012). In neurodegenerative diseases, Ca2+ signaling has been implicated in conditions as diverse as amyotrophic lateral sclerosis (Rowland and Shneider, 2001), Parkinson's disease (Thomas and Beal, 2007), skeletal muscle atrophy (Jackman and Kandarian, 2004), familial frontotemporal dementia (Furukawa ), Alzheimer's disease (Bezprozvanny and Mattson, 2008), scrapie (Kristensson ), progressive supranuclear palsy (Zemaitaitis ), Tourette syndrome (Martino ), Huntington's disease (Bezprozvanny and Hayden, 2004) and Wolfram syndrome (Osman ). Further understanding of how these and other diseases are interlinked through Ca2+ signaling is important for understanding the evolution of genetic disorders and developing more efficient treatment strategies in the future.

Here, we present a map of genes involved in Ca2+ signaling and the human genetic disorders associated with these genes. We provide an online database (http://cagedb.uhlenlab.org) that is automatically updated to incorporate novel discoveries related to gene functions and genetic diseases. Clustering genetic disorders with regard to common Ca2+ genes reveals interesting and unforeseen connections between human diseases. The Ca2+ genome and its associated online database will be valuable resources that should help facilitate a better understanding of the connection between Ca2+ signaling genes and associated diseases.

2 Materials and methods

2.1 Software

The main script was written in Ruby version 2.2.2 (https://www.ruby-lang.org). Data files from the queries were saved in SQLite version 3.7.14.1. The CaGeDB software tool can be downloaded from http://cagedb.uhlenlab.org.

2.2 Databases

Six different databases (Gene Ontology, Ensembl, NCBI Gene, OMIM, Comparative Toxicogenomic Database and Medical Subject Headings Database (MeSH)) were used in this study. A Ruby script automatically executed each query, and the results were saved in an SQLite database. A description of the databases used in each step is provided below. The data presented in this work were updated on February 19, 2017.

2.3 Genes associated with Ca2+

Gene ontology (GO, http://geneontology.org (Ashburner )) terms associated with Ca2+ signaling were determined by searching the GO names for text-strings related to Ca2+. The resulting list of GO terms was manually filtered to ensure true Ca2+-associations. The evidence codes associated with each GO annotation were not used in this study.

Two gene databases, Ensembl (http://www.ensembl.org) and NCBI Gene (http://www.ncbi.nlm.nih.gov/gene), were queried to search for genes classified with a Ca2+ signaling-related GO term from our list. We applied both Ensembl version 87 and the human genome assembly GRCh38.p7. The GO term-based queries were merged first according to their HGNC (HUGO (Human Genome Organization) Gene Nomenclature Committee (HGNC, http://www.genenames.org)) gene symbol and then according to their NCBI ID using mapping between the two database IDs provided by Ensembl. All of the query data files were saved in an SQLite database.

2.4 Genetic disorders related to Ca2+

Human diseases associated with the Ca2+ genome were determined by querying the Comparative Toxicogenomic Database (CTD, http://ctdbase.org (Davis ), from the Department of Bioinformatics at the MDI Biological Laboratory, ME, U.S. and the Department of Biological Sciences at the North Carolina State University, NC, U.S.), which uses the online database Mendelian Inheritance in Man (OMIM, http://omim.org, McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, MD, U.S.). Gene IDs used to query diseases were unique genes hits associated with the curated calcium GO-term ontology list and therefore the results of the disease query are independent on how many GO annotations a gene ID contained. The terms of the genetic disorders obtained from the OMIM database via the CTD were thereafter categorized using the Medical Subject Headings (MeSH, http://www.nlm.nih.gov/mesh/) database (Davis ).

The Ca2+-related genes and their associated diseases were represented as an adjacency matrix, and the distances between diseases were calculated as Pearson correlations. The numbers of clusters were determinate empirically with the ConsensusCluster Plus R package (Wilkerson and Hayes, 2010). Cluster stability was calculated by multiscale bootstrapping resampling (Suzuki and Shimodaira, 2006). A term frequency analysis per cluster was performed after the removal of numbers, punctuation and the words disease(s), disorder(s), with, and, syndrome, type and cell. The heat map representation shows term frequency by cluster and filtered by a variance of 90%. The analysis was performed using R software.

2.5 Graphical presentations and statistical analyses

The sunburst diagrams, force-directed graphs and network plots were generated using Gephi 0.8.2. The regression analyses and correlation coefficients were calculated using MATLAB.

3 Results

3.1 Identification of Ca2+ genes

To map genes related to Ca2+ signaling, we took advantage of Gene Ontology (GO) terms, established by the Gene Ontology Consortium (Gene Ontology, 2008). The GO project has developed three structured ontologies that describe gene products in a species-independent manner in terms of their associated biological processes, cellular components and molecular functions. We generated a list of GO terms by searching the entire database for strings matching entries related to Ca2+ signaling. The list was then manually checked to filter out entries exclusively involving Ca2+ signaling, which resulted in a total of 241 GO terms (Supplementary Table S1). Among these GO terms, 161 belonged to biological processes, 65 belonged to molecular functions and 15 belonged to cellular components. The large bias toward biological processes reflects the diverse role of Ca2+ signaling in regulating a wide variety of cellular processes (Berridge ).

All of the GO term annotations have a specific evidence code that reflects the type of work or analysis described in the cited reference that associates the gene to the GO term. Approximately three quarters (72%) of the identified Ca2+ signaling genes were annotated automatically without human curation, i.e. evidence code ‘Inferred from Electronic Annotation (IEA)’. This is a relatively low proportion because 98% of all GO annotations are based on evidence code IEA (du Plessis ). The other evidence codes are more equally distributed among annotations from experimental data and non-experimental data, and exact definitions of all of the evidence codes can be found elsewhere (http://geneontology.org/page/guide-go-evidence-codes).

We identified all of the human genes associated with the list of 241 Ca2+ signaling GO terms by querying two different gene databases, Ensembl and NCBI Gene. The Ensembl database contained 1739 genes that matched our list of GO terms, whereas the NCBI Gene database contained 1575 genes. Merging the two queries together resulted in a list of 1805 unique genes involved in Ca2+ signaling. Among these genes, 1668 had well-defined gene symbols. The GO terms associated with the highest number of Ca2+ signaling genes are shown in Figure 1B, and the largest term ‘calcium ion binding’ contained 751 genes. Altogether, the top 50 GO terms accounted for 4242 genes, thereby indicating that one gene can be linked to several GO terms.

3.2 Mapping Ca2+ genes to human genetic disorders

Next, we sought to map human genetic disorders associated with the Ca2+ genome. For this purpose, we used the open Online Mendelian Inheritance in Man database (OMIM, http://omim.org) via the Comparative Toxicogenomic Database (CTD, http://ctdbase.org). Among the 1805 mapped Ca2+ genes, 912 genes were associated with one or several genetic disorders. Searching the OMIM database resulted in a total number of 5205 disease-gene associations, with 1470 different genetic disorders as a result of a direct mutation in a Ca2+ gene, demonstrating that one specific gene can be linked to multiple diseases (or subtypes of a given disease).

The number of diseases associated with Ca2+ genes followed a power distribution (Fig. 2A). Datasets that have a large variance, which is the case for power distributions, are typical for scale-free networks (Malmersjo ; Smedler ). For example, PTGS2 (prostaglandin-endoperoxide synthase 2, also known as COX2) and IL6 (interleukin 6) were connected to approximately 100 diseases each (Fig. 2A), whereas CACNA1B (calcium channel, voltage-dependent, N-type, alpha 1B subunit) and TPT1 (tumor protein, translationally controlled 1) were only connected to one disease. Interestingly, the Ca2+ genome contained 21 genes that accounted for one quarter of all the gene-disease associations.

Fig. 2

GO term categories of annotated Ca2+ genes and their diseases. (A) Log-log plot of the distribution of diseases to genes associated with Ca2+. (B) Categorizing Ca2+-related GO terms into Target genes and Regulatory genes. (C) Percentage of Ca2+-related GO term categories with respect to genes (left) and genetic disorders (right). (D) Bar chart of the 20 annotated GO terms associated with the most Ca2+ genes plus genetic disorders. (E) Ca2+-related GO terms plotted as the number of genes versus the number of diseases

To determine the specific role of individual genes involved in Ca2+ signaling, we divided the GO terms into two major classes: ‘Target genes’ and ‘Regulatory genes’ (Fig. 2B). ‘Target genes’ typically respond to changes in the intracellular Ca2+ concentration, such as kinases and transcription factors. ‘Regulatory genes’ are defined as genes that regulate the intracellular Ca2+ concentration, such as Ca2+ channels, Ca2+ pumps and Ca2+ binding proteins (including Ca2+ buffer proteins). ‘Regulatory genes’ were further classified with regards to their regulatory function, where increases in free cytosolic Ca2+ are termed ‘ON-reactions’ and decreases are termed ‘OFF-reactions’ as described by Michael Berridge (Berridge ). We also classified the GO terms belonging to ‘Regulatory genes’ according to their cellular location (Fig. 2B). The schematic diagram in Figure 1A shows the typical proteins encoded by ‘Regulatory genes’ that are responsible for both ‘ON-reactions’ and ‘OFF-reactions’.

Interestingly, when the Ca2+-related diseases were divided into our two main categories, we found that ‘Regulatory genes’ accounted for more than 70% of the total number of diseases (Fig. 2C), which demonstrates that perturbations in the Ca2+ signaling machinery can generate a large number of genetic disorders. Furthermore, one quarter of all of the gene-disease associations involved genes related to increases in cytosolic Ca2+ concentrations. The 20 GO-terms with the largest number of diseases are listed in Figure 2D, with ‘calcium ion binding’ number one. When the entire list of GO terms was plotted using the number of genes and the number of diseases as axes, the terms appeared correlated (Fig. 2E). Performing a linear regression analysis showed a strong correlation (R = 0.95707) among the GO terms.

3.3 Analyzing diseases linked to Ca2+ signaling

We then matched the genetic disorders queried from the OMIM database to the Medical Subject Headings (MeSH), which is the National Library of Medicine's controlled vocabulary thesaurus. MeSH descriptors are organized in 16 categories, with Diseases comprising one category. Each category is further divided into subcategories, in which descriptors are arrayed hierarchically from most general to most specific for up to 13 hierarchical levels. For example the subcategory ‘Pathological Conditions, Signs, and Symptoms’ has a scope ‘Abnormal anatomical or physiological conditions and objective or subjective manifestations of disease, not classified as disease or syndrome’ with up to 793 associated descriptors, i.e. symptoms associated with edema, heart murmurs, renal colic, fatigue, etc. (For a full description of MeSH vocabulary go to the MeSH website: https://www.nlm.nih.gov/mesh). The structure of the branching hierarchy frequently represents a compromise among the views and needs of particular disciplines and users, in the absence of any single universally accepted arrangement. As such, each MeSH descriptor appears in at least one place, and may appear in as many additional places as may be appropriate (The complexity of the branching is apparent in Figure 3A and Supplementary Fig. S1). Out of the entire list of 11 721 MeSH terms describing human diseases, 13% (1470 terms) had at least one reported association with a gene related to Ca2+ signaling (Fig. 3A). The category of ‘neoplasms (C04)’ i.e. abnormal growth in a body, such as tumors, was the most common main category, and 975 gene connections were observed between the category and its sub-categories (Fig. 3B). Remarkably, Ca2+ associated diseases encompass at least one sub-category in every branch of the MeSH tree hierarchy, except for the parent category ‘Disorders caused by external forces rather than by physiologic dysfunction or by pathogens (C21)’. The sub-category with the most associated genes was ‘prostatic neoplasms’ (prostate cancer), which presented 74 gene connections.

Fig. 3

Genetic disorders associated with Ca2+ genes. (A) Sunburst diagram of the MeSH term hierarchy. Fields/terms with at least one Ca2+ gene are labeled red, and those without associations are labeled blue. The sector angles correspond to the number of children each term includes. Sub-terms are further away from the center than parent terms. (B) Bar chart of the MeSH parent terms (inner circle of panel A) and the cumulative associated Ca2+ genes. (C) Heat map of Pearson’s correlation values of diseases and hierarchical clustering of groups. Each color shows clusters selected by consensus clustering of k = 5. (D) Disease names analyzed for term frequency of each cluster, filtered for the top 10% of terms with the highest variance

When the relationships between GO terms and gene diseases were plotted with a force-directed graph (Supplementary Fig. S1), properties of scale-free network became apparent. Several large nodes originating from GO terms (pink dots) were observed with more than 500 genes (yellow dots) for neoplasms (MeSH ID C04), cardiovascular diseases (MeSH ID C14), digestive system diseases (MeSH ID C06), nervous system diseases (MeSH ID C10) and pathological conditions, signs and symptoms (MeSH ID C23). Similarly, the number of diseases (red dots in Supplementary Fig. S1) connected to each gene varied greatly.

The Comparative Toxicogenomic Database disease vocabulary (http://ctdbase.org) uses modified descriptors from MeSH combined with genetic disorders from OMIM to map OMIM diseases within the hierarchical MeSH disease vocabulary, thus expanding disease representation. When organizing the top five Ca2+ genes based on the number of associated diseases using the CTD database, IL6 (Interleukin 6) and PTGS2/COX2 were the top hits, with 108 and 105 MeSH terms respectively, followed by TGFB1 (transforming growth factor, beta 1) with 76 MeSH terms, and NOS2 (nitric oxide synthase 2) and AGT (Angiotensinogen) with 69 and 64 MeSH terms, respectively. Common to most of these genes is their key role in inflammatory responses, which underlies a myriad of diseases. For example, IL6 is produced at the site of inflammation, playing a key role in the acute phase response (Tanaka ), PTGS2/COX2 in the production of inflammatory prostaglandins from arachidonic acid (Alhouayek and Muccioli, 2014). Additionally NOS2, as well as being a messenger molecule with diverse functions, mediates cysteine S-nitrosylation of PTGS2/COX2 and enhances the synthesis of IL6. Interestingly, NOS2 transcription is modulated by the Ca2+ dependent transcription factor NF-κB (nuclear Factor kappa-light-chain-enhancer of activated B cells) (Zhou ). TGFB1, as well as functioning as a proinflammatory cytokine, plays an important role in bone remodeling, a pivotal role in tumor suppression (and paradoxically tumor promotion) and can induce epithelial-to-mesenchymal transition (Hwangbo ). Whereas AGT is an essential component of the renin-angiotensin system, being a potent regulator of blood pressure, body fluid and electrolyte homeostasis.

We also sought to determine whether diseases could be grouped together in an unbiased manner with regards to their association with Ca2+ signaling genes to reveal relationships between diseases with no apparent connections through the clustering of genes involved in Ca2+ signaling. Hence, diseases were clustered based on their number of shared Ca2+ genes. Generating a correlation matrix based on the entire Ca2+ genome yielded five major groups of clustered diseases (Fig. 3C). We hypothesized that the clusters corresponded to diseases with shared etiologies or mechanisms of action for the pathology itself. Interesting relations between diseases occurred when the disease names were analyzed by cluster (Fig. 3D). For example, Cluster 1 was rather homogeneous because it mainly contained developmental disorders, such as muscular dystrophy, hereditary nephritis (Alport Syndrome), and Alveolar capillary dysplasia, whereas Cluster 2 was more heterogeneous and included diverse diseases, such as multiple sclerosis, diabetes mellitus, chronic kidney failure and pulmonary fibrosis. Cluster 5 contained mostly cancer related diseases, such as skin neoplasms, B-cell lymphoma and breast carcinoma. Additionally, Cluster 5 also contained asthma and amyotrophic lateral sclerosis.

3.4 CACNA1C-associated gene network

Finally, we wanted to perform an analysis of one Ca2+ gene that is known to be involved in several diseases, preferably in different organs. Thus, we searched our database for genetic disorders associated with CACNA1C, which encodes the alpha-1 subunit of the voltage-dependent Ca2+ channel Cav1.2. Eight genetic disorders with mutations in the CACNA1C gene were found: cardiac arrhythmia, autistic disorder, bipolar disorder, brugada syndrome 3, hypertension, hypoglycemia, immunologic deficiency syndromes, congenital limb deformities and Timothy syndrome. These diseases are mainly caused by mutations in CACNA1C that occur during embryonic or fetal development, although the effects may not be observed until later in life.

Plotting a network map that was expanded one degree revealed 67 genes linked to CACNA1C-related diseases (Fig. 4). Of the 8 associated diseases, four (timothy syndrome, Brugada Syndrome 3, Limb Deformities, congenital) have CACNA1C as their sole Ca2+-related gene involved, whilst the other remaining diseases had one or more Ca2+-related genes linked with the disease. The two CACNA1C-associated neurological disease families, autistic disorder and bipolar disorder, of which both had the highest numbers of associated calcium-related genes, shared the following genes: PDE4B, NTRK2, GRIK2, DRD1 and HTR2A. Enrichment analysis of the GO annotations linked to these particular genes revealed that they are involved in a range of biological processes including the modulation of synaptic transmission, synaptic plasticity, blood circulation, muscle contraction and regulation of actin filament based movements (Fig. 4B). Comparing shared calcium related genes between cardiac arrhythmias and autistic disorder (PTGS2, GJA1 and AVP), the enrichment analysis of the GO ontology revealed biological processes ranging from the regulation of system and homeostatic processes, vasoconstriction, renal system processes, cell communication by electrical coupling and regulation of lipid biosynthesis (Fig. 4C).

Fig. 4

CACNA1C disease network with associated Ca2+ genes. (A) Network of disease categories (blue) linked to CACNA1C (red) expanded one degree, which indicated 67 associated Ca2+ genes (green). In (B) and (C), dot plot displaying the top 20 GO terms from the GO-enrichment analysis of the genes common to: (B) Autistic Disorder and Bipolar Disorder, (C) Autistic Disorder and Cardiac Arrhythmias. Color bar code display the adjusted P-values for multiple comparisons. Dot sizes display number of associated genes

The analysis of the interrelationships among the Ca2+ signaling genes and their associated diseases will not only increase our general understanding of how genetic disorders are connected and develop but also open up new approaches for treating these common diseases.

4 Discussion

Our results demonstrate the key role of Ca2+ in regulating important cell processes in virtually all of the different cell types of the human body. First, all of the genes involved in intracellular Ca2+ signaling in human cells were mapped. Second, human genetic disorders with reported mutations in the genes associated with Ca2+ signaling were determined. Third, an automatically updated online research tool (http://cagedb.uhlenlab.org) for identifying genes and diseases linked to Ca2+ signaling was established.

Central to this study are the GO terms, which facilitate the identification of genes related to Ca2+ signaling. New GO terms are constantly established as a result of novel scientific discoveries of previously annotated genes (Blake, 2013). By applying text-mining algorithms to scientific publications, new GO term annotations can be determined (Bayes ). Here, we used only the assigned annotations. A quarter of all of the annotations of our genes were based on experimental data, which indicates reliability (Skunca ). The set of GO terms used in the online research tool implemented here is manually updated and currently consists of 241 terms.

Our unbiased clustering of diseases based on genes involved in Ca2+ signaling, which is presented in Figure 3C, D, is an example of how novel hypothesizes can be generated using this online research tool. The results of this analysis can be investigated further by determining the overlapping genes between diseases with no apparent connections. For example, our analysis revealed connections between amyotrophic lateral sclerosis and various cancers. Such information can, for instance, be used to test if a successful treatment for a certain disease is applicable for another disease because of their overlapping Ca2+ signaling genes. Moreover, the gene network for CACNA1C presented in Figure 4 is another example of how this tool can be used. This network showed that the genes PTGS2, GJA1 and AVP as well as CACNA1C were all linked to the two different organ disease groups autistic disorder and cardiac arrhythmias. The PTGS2 gene is linked to the greatest number of diseases. GJA1 (Gap junction alpha-1) is also known as Cx43 (connexin 43), a protein that enables the conduction of Ca2+ between cells and the extracellular space via gap junctions and hemichannels. AVP (arginine vasopressin) is a posterior pituitary hormone that has antidiuretic effects on the kidney and can also contract smooth muscles during parturition and lactation. AVP is also involved in cognition, tolerance, adaptation and complex sexual and maternal behavior. Additionally, the network analysis of CACNA1C related diseases showed that the shared genes between the two psychiatric disorders were mainly involved in synaptic transmission, blood circulation and muscle contraction. Interestingly, evaluation of the excitability and viability of central motor pathways of the human motor cortex by transcranial magnetic stimulation has been widely used for the investigation of a variety of neurologic and psychiatric disorders (Chroni ). Additionally, changes observed in autism include alterations in the signaling pathways mediating neurovascular coupling as a result of an increase in synaptic inhibition (Reynell and Harris, 2013). Cerebral blood flow and cerebral metabolic rate are closely correlated and are conceptualized as proxies of synaptic transmission with a number of developmental, degenerative, neoplastic and ischemic processes associated with an uncoupling of blood flow and metabolism (Ota ).

During the identification of the Ca2+ genes and their related diseases for this study, the strength of the different associations varied considerably. The strength (or evidence) of a certain gene-disease association was based on a number of methods and standards for measuring, validating and interpreting genetic associations (Khoury ). Additionally, the selection criteria used to identify the genes involved in Ca2+ signaling were not always entirely reliable because the Gene Ontology database does not contain this specific category. However, we decided to risk the inclusion of false positives, i.e. gene-disease associations that are not actually related to Ca2+ signaling, rather than miss true gene-disease associations.

Annotations are either automatically generated or manually added to databases by curators. These processes can certainly be further improved, which would result in more specific and up-to-date annotations (du Plessis ). Currently, scientists must report their own findings to be absolutely sure that the results are annotated accurately in the associated databases. The purpose of this work was to survey available databases to identify genes related to Ca2+ signaling and their associated diseases. As a consequence, a number of Ca2+ gene-disease associations were most likely not included because they have not yet been annotated. In the future, further improvements to machine learning will increase the reliability of the annotation process.

In summary, we have mapped 1805 genes and 1470 human genetic disorders that are involved in Ca2+ signaling. Our online research tool (http://cagedb.uhlenlab.org) is a dynamic database that is updated weekly with new genes and genetic disorders associated with Ca2+ signaling. This online database will be a valuable resource that should help facilitate a better understanding of the network interactions between Ca2+ signaling genes and diseases. Our study confirms the important role of Ca2+ in living cells and shows that perturbed Ca2+ handling can result in a large number of diverse human diseases.

Click here for additional data file.